EXTANT SEED PLANTS

Plant woody, evergreen; nicotinic acid metabolised to trigonelline; primary cell walls rich in xyloglucans and/or glucomannans, 25-30% pectin [Type I walls]; lignins rich in guaiacyl units; true roots present, xylem exarch; shoot apical meristem complex; arbuscular mycorrhizae +; stem with ectophloic eustele, endodermis 0, xylem endarch; vascular tissue in t.s. discontinuous by interfascicular regions; vascular cambium + [xylem ("wood") differentiating internally, phloem externally]; wood homoxylous, tracheids +; tracheid/tracheid pits circular, bordered; sieve tube/cell plastids with starch grains; phloem fibers +; stem cork cambium superficial, root cork cambium deep seated; nodes ?; leaf vascular bundles collateral; leaves spiral, simple, axillary buds?, prophylls [including bracteoles] two, lateral; plant heterosporous, sporangia eusporangiate, on sporophylls, sporophylls aggregated in indeterminate cones/strobili; true pollen [microspores] +, mono[ana]sulcate, pollen exine and intine homogeneous, ovules unitegmic, crassinucellate, megaspore tetrad tetrahedral, only one megaspore develops, megasporangium indehiscent; male gametophyte development endo/exosporic, gametes two, with cell walls; female gametophyte endosporic, initially syncytial, walls then surrounding individual nuclei; seeds "large", first cell wall of zygote transverse, embryo straight, endoscopic [suspensor +], short-minute, with morphological dormancy, white, cotyledons 2; plastid transmission maternal; two copies of LEAFY gene, PHY gene duplication, mitochondrial nad1 intron 2 and coxIIi3 intron present.

MAGNOLIOPHYTA

Plant woody, evergreen; lignans, O-methyl flavonols, dihydroflavonols, triterpenoid oleanane, non-hydrolysable tannins, quercetin and/or kaempferol +, apigenin and/or luteolin scattered, cyanogenesis via tyrosine pathway, lignins derived from both coniferyl and sinapyl alcohols, containing syringaldehyde [in positive Maüle reaction, syringyl:guaiacyl ratio less than 2-2.5:1], and hemicelluloses as xyloglucans; root apical meristem intermediate-open; root vascular tissue oligarch [di- to pentarch], lateral roots arise opposite or immediately to the side of [when diarch] xylem poles; origin of epidermis with no clear pattern [probably from inner layer of root cap], trichoblasts [differentiated root hair-forming cells] 0; stem with 2-layered tunica-corpus construction; wood fibers and wood parenchyma +; reaction wood ?, with gelatinous fibres; starch grains simple; primary cell wall mostly with pectic polysaccharides; tracheids +; sieve tubes eunucleate, with sieve plate, companion cells from same mother cell that gave rise to the tube, the sieve tube with P-proteins; nodes unilacunar; stomata with ends of guard cells level with aperture, paracytic; leaves with petiole and lamina [the latter formed from the primordial leaf apex], development of venation acropetal, 2ndary veins pinnate, fine venation reticulate, vein endings free; flowers perfect, polysymmetric, parts spiral [esp. the A], free, numbers unstable, P not differentiated, outer members not enclosing the rest of the bud, A many, development centripetal, with a single trace, introrse, filaments stout, anther ± embedded in the filament, tetrasporangiate, dithecal, with at least outer secondary parietal cells dividing, each theca dehiscing longitudinally by action of hypodermal endothecium, endothecial cells elongated at right angles to long axis of anther, tapetum glandular, binucleate, microspore mother cells in a block, microsporogenesis successive, pollen subspherical, binucleate at dispersal, trinucleate eventually, tectum continuous, endexine compact, lamellate only in the apertural regions, pollen tube elongated, with callose plugs, penetrating between cells, growth rate moderate, siphonogamy occuring, nectary 0, G free, several, ascidiate, with postgenital occlusion by secretion, few [?1] ovules/carpel, ovules marginal, anatropous, bitegmic, micropyle endostomal, integuments 2-3 cells thick, megasporocyte single, megaspore lacking sporopollenin and cuticle, chalazal, female gametophyte ?type, stylulus short, stigma ± decurrent, wet [secretory]; P deciduous in fruit; seed exotestal; double fertilisation +, endosperm ?diploid, cellular [first division oblique, micropylar end initially with a single large cell, chalazal end more actively dividing], copious, oily and/or proteinaceous, embryo cellular ab initio; germination hypogeal, seedlings/young plants sympodial; Arabidopsis-type telomeres [(TTTAGGG)_n]; whole genome duplication, single copy of LEAFY and RPB2 gene, knox genes extensively duplicated [A1-A4], AP1/FUL gene, paleo AP3 and PI genes [paralogous B-class genes] +, with "DEAER" motif, SEP3/LOFSEP and PHYA/PHYC gene pairs.

Possible apomorphies are in bold. Note that the actual level to which many of these features, particularly the more cryptic ones, should be assigned is unclear, because some taxa basal to the [magnoliid + monocot + eudicot] group have been surprisingly little studied. Furthermore, details of relationships among gymnosperms will affect the level at which some of these characters are pegged.

NYMPHAEALES [AUSTROBAILEYALES [[CHLORANTHALES + MAGNOLIIDS] [MONOCOTS [CERATOPHYLLALES + EUDICOTS]]]]: vessels +, elements with scalariform perforation plates; pollen tectate-columellate, tectum reticulate [perforated]; nucleus of egg cell sister to one of the polar nuclei; ?genome duplication; "DEAER" motif in AP3 and PI genes lost, gaps in these genes.

PROTEALES [TROCHODENDRALES [BUXALES [GUNNERALES + CORE EUDICOTS]]]: ?

TROCHODENDRALES [BUXALES [GUNNERALES + CORE EUDICOTS]]: benzylisoquinoline alkaloids 0; euAP3 + TM6 genes [duplication of paleoAP3 gene: B class], mitochondrial rps2 gene lost.

BUXALES [GUNNERALES + CORE EUDICOTS]: ?

GUNNERALES + CORE EUDICOTS: Ellagic and gallic acids common, cyanogenesis via phenylalanine, isoleucine or valine pathways; micropyle?; PI-dB motif +, small deletion in the 18S ribosomal DNA common.

CORE EUDICOTS: Root apical meristem closed; flowers rather stereotyped: 5-merous, parts whorled, K and C distinct, K with 3 traces, A = 2x K, (often numerous, but then often fasciculate and/or centrifugal), pollen tricolporate, (nectary disc +), [G 5], [3] also common, compitum +, placentation axile, stigma not decurrent; endosperm nuclear; fruit dry, dehiscent, loculicidal [when a capsule]; euAP1 + euFUL + AGL79 genes [duplication of AP1/FUL or FUL-like gene], PLE + euAG [duplication of AG-like gene: C class], SEP1 + FBP6 genes [duplication of AGL2/3/4 gene].

SAXIFRAGALES [VITALES + ROSIDS]: Stipules +.

VITALES + ROSIDS: Anthers articulated [± dorsifixed, transition to filament narrow, connective thin].

ROSIDS: embryo long; genome duplication; chloroplast infA gene defunct, mitochondrial coxII.i3 intron 0.

ROSID I: Endosperm scanty.

CELASTRALES [OXALIDALES + MALPIGHIALES]: seed exotegmic, cells fibrous.

OXALIDALES + MALPIGHIALES: ?

MALPIGHIALES Martius  Main Tree, Synapomorphies.

Vessel element type?; (sieve tubes with non-dispersive protein bodies); leaf margin toothed [teeth with a single vein running into a congested ± deciduous apex], stipules +; stigma dry; exotegmen fibrous. - 39 families, 716 genera, 15935 species.

Crown Malpighiales probably diverged some time in the Cretaceous-late Aptian, perhaps some 114 million years before present ([119.4-]113.8[-110.7]/[105.9-]101.6[-101.1] million years before present - high and low estimates: Davis et al. 2005a); diversification seems to have been rapid (Wikström et al. 2001, however, suggest divergence dates of 91-88 million years before present, diversification beginning only 81-77 million years before present, with stem groups of many families being evident before the beginning of the Tertiary). The order contains ca 7.8% eudicot diversity (Magallón et al. 1999) and is particularly important in tropical rainforests where it is an important component of the diversity of the understory; it accounts for up to some 28% of the species and 38% of the total stems (Davis et al. 2005a); members of Ericales are the other main component.

The butterfly Cymothoë has hosts widely scattered in this order (Ackery 1988), although also found on Bignoniaceae (one species) and Rhamnaceae (sometimes another species). Phyllonorycter leaf-mining moths (Lepidoptera -Gracillariidae - Phyllocnistinae) seem to have diversified on this clade (and especially Fagales) some time in the region of 50.8-27.3 million years before present, well after the Malpighiales diversified, and after the genus itself evolved, some 76.3-50.3 million years before present (Lopez-Vaamonde et al. 2006).

Articulation of the pedicels is another feature that may be common to the order.

Although Malpighiales are strongly supported as being monophyletic (e.g. Davis et al. 2005a), relationships within them are still poorly understood (e.g. Soltis et al. 2007a). Those in the tree shown here are mostly taken from Litt and Chase (1999), Schwarzbach and Ricklefs (2000), Chase et al. (2002), and Davis and Chase (2004), but they are in general agreement with relationships apparent in broader studies. Davis et al. (2005a) have recently clarified some relationships in Malpighiales in a four-gene (all three compartments) analysis, in particular suggesting an association between the families with parietal placentation (and also Goupiaceae) and that Centroplacus should be recognised as a separate family (see also Korotkova et al. 2007 for relationships in Malpighiales). The inclusion of Rafflesiaceae in Malpighiales follows the recent findings of Barkman et al. (2004, 2007), Davis and Wurdack (2004), and in particular Davis et al. (2006), who place it with strong support as sister to Euphorbiaceae s. str. It seems useful to adopt a narrow circumscription for families that used to be included in Flacourtiaceae and Euphorbiaceae s.l. Even if future work suggests reaggregation of genera that used to be placed in these two families, groupings within these new units will be different from those suggested by previous classifications. Note that the realignments caused by the break-up of the old Flacourtiaceae and integration with Salicaceae and Achariaceae correlate well with a number of morphological/anatomical characters. Paracytic stomata may characterise a sizeable clade in Malpighiales...

See Endress and Matthews (2006b) for petal appendages, etc., in the order; Tokuoka and Tobe (2006) integrate testa and anatomy with phylogeny.

Includes Achariaceae, Balanopaceae, Bhesa, Bonnetiaceae, Caryocaraceae, Centroplacaceae, Chrysobalanaceae, Clusiaceae, Ctenolophonaceae, Dichapetalaceae, Elatinaceae, Erythroxylaceae, Euphorbiaceae, Euphroniaceae, Goupiaceae, Humiriaceae, Hypericaceae, Irvingiaceae, Ixonanthaceae, Lacistemataceae, Linaceae, Lophopyxidaceae, Malpighiaceae, Malesherbiaceae, Medusagynaceae, Ochnaceae, Pandaceae, Passifloraceae, Peraceae, Phyllanthaceae, Picrodendraceae, Podostemaceae, Putranjivaceae, Quiinaceae, Rafflesiaceae, Rhizophoraceae, Salicaceae, Trigoniaceae, Turneraceae, Violaceae.

Synonymy: Balanopales Engler, Chrysobalanales Reveal & Doweld, Elatinales Nakai, Euphorbiales Lindley, Flacourtiales Heinze, Homaliales Bromhead, Huales Doweld, Hypericales Dumortier, Irvingiales Doweld, Lacistematales Baskerville, Linales Baskerville, Rafflesiales Oliver, Marathrales Dumortier, Medusagynales Reveal & Doweld, Ochnales Reveal, Pandales Engler & Gilg, Passiflorales Dumortier, Phyllanthales Doweld, Podostemales Lindley, Rhizophorales (Persoon) Reveal & Doweld, Salicales Lindley, Samydales Dumortier, Scyphostegiales Croizat, Turnerales Dumortier, Violales Perleb - Euphorbianae Reveal, Ochnanae Doweld, Podostemanae Reveal, Rafflesianae Reveal, Rhizophoranae Reveal & Doweld, Violanae Reveal - Malpighiopsida Bartling, Passifloropsida Brongniart, Podostemopsida G. Cusset & C. Cusset, Salicopsida Bartling, Violopsida Brongniart

Achariaceae + Violaceae + Salicaceae + Lacistemataceae + Goupiaceae + Malesherbiaceae + Turneraceae + Passifloraceae: sieve tubes with non-dispersive protein bodies; cuticle waxes usu. 0; (foliar glands +); pedicels articulated; nectariferous tissue, A = and opposite K, G with median member abaxial, placentation parietal; seed arillate; endotegmen persistent; endosperm persistent, oily.

All six major clades in this group (Malesherbiaceae + Turneraceae + Passifloraceae as a single clade) may have diverged in the Cretaceous-Albian 111-100 million years before present, or somewhat later (Davis et al. 2005a, details are given for the individual clades: Wikström et al. 2001 suggests many of the clades do not diverge until [well] after 63 million years before present). Larvae of butterflies such as Nymphalidae-Acraeinae and N.-Nymphalinae-Heliconiini, -Vagrantini and -Argynnini commonly eat members of this group (Ehrlich & Raven 1964; see also Dahlgren & van Wyk 1988; Arbo 2006; Simonsen 2006; Silva-Brandão et al. 2008; etc.); this is also discussed under individual families below. Some Acraeinae in particular may cue on the presence of the cyanogenic glucoside gynocardin in potnatial food plants, indeed, that larvae of Acraea horta, normally living on the woody Kiggelaria africana, ate herbaceous Achariaceae prompted the successfull search for that compound in the latter family (Steyn et al. 2002). Toxic compounds like gynocardin may be sequestered by the larva and passed on to the adult.

There are numerous anatomical, chemical and floral links between Salicaceae, Achariaceae and Violaceae, sometimes also Passifloraceae, Malesherbiaceae and Turneraceae (Nandi et al. 1998). Thus Achariaceae + Malesherbiaceae + Turneraceae + Passifloraceae have in common cyclopentenoid cyanogenic glucosides and/or cyclopentenyl fatty acids and also commonly some sort of corona or scales on petals. Molecular evidence that the whole group formed by families with parietal placentation is monophyletic has previously not been compelling (e.g. see Savolainen et al. 2000a; Chase et al. 2002), although part of the rpS 16 gene is absent from Passifloraceae, Violaceae, Salicaceae s. str., and Turneraceae (and also Linaceae and Malpighiaceae, so really a feature of Malpighiales?: see Downie & Palmer 1992). Salicaceae are weakly associated with Passifloraceae, and in turn with Humiriaceae and Pandaceae, in Chase et al (2002). Associations like Salicaceae + Passsifloraceae s.l. (inc. Turneraceae and Malesherbiaceae) are weakly supported, as are Achariaceae + Violaceae (and Lacistemataceae + Ctenolophonaceae)- see below. Tokuoka and Tobe (2006) found a weakly-supported relationship between the Passifloraceae group and Violaceae (see also Soltis et al. 2007a), and strongly supported relationships between Lacistemataceae and Salicaceae. However, Davis et al. (2005a) found a moderately supported association of these taxa with parietal placentation (59% bootstrap, 100% posterior probability), and also Goupiaceae, with axile placentation, and a similar grouping is also evident in e.g. Wikström et al. (2001) and Korotkova et al. (2007: 83% jacknife, 1.00 pp, Goupiaceae not included). Ixonanthes was rather surprisingly embedded in Achariaceae in the Bayesian analysis of Soltis et al. (2007a).

There is much information on seed anatomy in Takhtajan (1992) while Krosnick et al. (2006) briefly discuss the evolution of polyandry in this group - in some cases, at least, the numerous stamens form a single whorl.

ACHARIACEAE Harms, nom. cons.   Back to Malpighiales

Shrubs to trees ((climbing) herbs); cyclopentenoid cyanogenic glucosides and/or cyclopentenyl fatty acids +, gynocardin, ellagic acid [Kiggelaria] +; vessel elements with simple or scalariform perforations; fibers septate; petiole bundle annular, with two lateral/adaxial strands (inverted medullary plate - Lindackeria); leaves spiral or 2-ranked, margins entire (serrate; stipules 0), petiole often geniculate; (plant dioecious); inflorescence spicate or cymose (fasciculate); K and C spiral or not, not in a simple alternating relationship, K 2-5, C 4-15 (3-4, connate - Acharieae), often in two series, (adaxial scales +), A 5-many, opposite C or irregular, centripetal or simultaneous, (from a ring meristem), anthers basifixed, elongate (barely so - Chiangodendron; porose; locellate), pollen also tricolporoidate, disc 0, G [2-10], median member?, ovules sessile, (atropous - Xylotheca, Hydnocarpus, Lindackeria; 1/carpel), (archesporium bicellular - Caloncoba) (embryo sac Allium-type [bisporic, 8-nucleate] - Acharieae), integuments 5-6 cell layers across, micropyle endo- or bistomal or zigzag, nucellar cap and epistase +, style (short), branched or not, stigma capitate-peltate; fruit also a berry; seed (arillate), coat thick, pachychalazal [not Hydnocarpus laurifolius], or both integuments multiplicative [possibly simply different interpretations], testa vascularised, sarcotestal and with stomata (Acharieae) or not, (inner mesotesta sclereidal), endotesta lignified, cells sclereidal (radially elongated); endosperm copious, suspensor 0, embryo green; n = 10, 12, 23.

30[list]/145: Hydnocarpus (40). Pantropical. (Map: from Sleumer 1954, 1980; Şerban Procheş, pers. comm. [Africa]; Andrew Ford, pers. comm. [Australia])[Photo - Flower, Fruit, Fruit, Acharia tragodes - Leaves.]

• Achariaceae can be recognised by their stipulate leaves, often entire leaf blades and pulvinate petioles, and flowers in which the number of sepals and petals is often clearly not the same, and/or the perianth is not simply biseriate, and/or the petals have basal, adaxial scales. There is no nectary, and the anthers are usually long. The fruits are often quite large, the placentation is parietal, and the seeds often have conspicuous vascular bundles in the coat.

The feeding behaviour of Acraeini butterfly larvae are consistent with the expanded family limits adopted here (Steyn et al. 2002, 2003 and references).

Circumscription (see Chase et al. 2002; Sosa et al. 2003): Acharieae (more or less herbaceous and viny; no testal bundles; zig-zag micropyle; fibrous exotegmen), Erythrospermeae (Erythrospermum - fibrous exotegmen), Pangieae (inc. Kiggelarieae) and Lindackerieae (Oncobeae minus Oncoba). Note that the family is divided into three strongly-supported clades, largely Hydnocarpus, Erythrospermeae + Lindackerieae, and Acharieae + Pangieae, and support for monophyly of the family as a whole is strong (Sosa et al. 2003).

Gynocardin is rarely found in other families of this group. Cyanogenic glycosides may be absent. There are large and medium intervascular pits; the wood also has solitary pores and lacks tracheids. Lindackeria has superficial cork cambium. Pollen variation is considerable (Wendt 1988).

The bulk of Achariaceae have almost universally been included in Flacourtiaceae s.l. (Cronquist 1981; Takhtajan 1997) = Salicaceae, q.v. for relationships.

Information is taken from Hegnauer (1966, 1989, as Flacourtiaceae: chemistry), van Heel (1977, 1979: testa anatomy), Endress and Voser (1975: floral development), Miller (1975: wood anatomy), Spencer and Seigler (1985: chemistry), Lemke (1988), Gavrilova (1998: pollen) and Steyn et al. (2002a, b, 2003: ovule development, testa anatomy); see also Judd (1997a) and especially Chase et al. (2002). Bernhard and Endress (1999) discuss androecial initiation. I thank Sue Zmarzty for comments.

Synonymy: Erythrospermaceae Doweld, Kiggelariaceae Link, Pangiaceae Endlicher

VIOLACEAE Batsch, nom. cons.   Back to Malpighiales

Herbs to trees (lianes); plants often Al accumulators; tannins 0 [woody members?]; vessel elements with simple or scalariform perforations; calcium oxalate often as crystals; petiole bundles arcuate; stomata para- or anisocytic; leaves spiral or 2-ranked (opposite), margins involute, colleters +, (stipules petiolar; lobed); pedicels articulated; flowers poly- or monosymmetric, K quincuncial, abaxial C spurred or not, A (3), all or 2 abaxial with abaxial nectary, (filaments connate), anthers connivent, connectives prolonged (± 0), G [(2) 3(-5)], micropyle zigzag (endostomal), hypostase +, styles separate or style +, stigmatic head subcapitate, asymmetrical or not, receptive area small; (fruit a berry, nutlike), K often persistent; seeds (winged) often arillate/carunculate, exotesta subpalisade to tabular, ± thickened, (mesotesta sclerenchymatous), endotesta usu. crystalliferous; embryo (small), green [Viola]; n = 6-13+.

23[list]/800: Viola (400-600: cleistogamy widespread, V. tricolor, the pansy, and the related V. arvensis important in early studies of genetics and speciation), Rinorea (160-270), Hybanthus (90-150). World-wide; woody taxa esp. in the lowland tropics (Map: from Hultén 1958, 1971; George 1982; Hultén & Fries 1986; Hekking 1988 - incomplete for South America). [Photo - Leonia, Alexis fruit and flowers, Viola.]

• Violaceae may be recognised by their serrate, stipulate leaves which often dry yellowish and with raised reticulum, their flowers, which are often spurred and have stamens with short filaments and very well-developed connectives that closely surround the gynoecium, and their fruit, which is often a distinctive loculicidal capsule, the rather large seeds being dispersed explosively as the fruit wall closes round them and squeezes them out. Plants are herbs, trees or lianes.

Violaceae are the preferred hosts of the majority of fritillaries (Nymphalidae: Argynnini - see Simonsen 2006).

Melicytus is woody, dioecious, and baccate; the flowers are almost radially symmetrical... Not surprisingly, woody Violaceae are quite commonly often wrongly identified or not named at all, as Ron Leisner will confirm - however, the vegetative characters mentioned above do help. Viola has storied cambium. Fusispermum has a contorted corolla. In Anchietea and Decorsella the seeds mature exposed on the open carpels.

Violaceae are weakly associated with Achariaceae (and Goupiaceae, Lacistemataceae and Ctenolophonaceae) in Chase et al (2002a).

For embryology, etc., see Singh (1970), for chemistry, see Hegnauer (1970, 1990), for general information, see Hekking (1988) and Munzinger and Ballard (2003: also key to genera, two undescribed).

Synonymy: Alsodeiaceae J. Agardh, Leoniaceae A. L. de Candolle

SALICACEAE Mirbel, nom. cons.   Back to Malpighiales

(Ectomycorrhizal) evergreen (deciduous) trees; cocarcinogen, (ellagic acid +), tanniniferous; cork?; vessel elements with simple or scalariform perforations; (nodes 2:2 - some Azara); petiole bundle arcuate or annular with flange bundles; stomata ?; leaves spiral or 2-ranked (opposite), supervolute-curved or involute, teeth (margin entire; venation palmate; glands +; stipules 0); (plant dioecious), inflorescence various; flowers often small, 3-6-merous, (hypanthium +), K (0-)3-8(-15), often valvate (connate), C 0 or = K (more than K - e.g. some Scolopieae; basally connate; corona +), disc often with glands or lobes (the latter intrastaminal; 0), A 1 to many (fasciculate, opposite C), centrifugal (simultaneous) initiation, anthers (extrorse), ellipsoid to subglobose (linear), (pollen inaperturate - Populus), G [2-5(-13)], (to inferior), (placentation axile; basal), ovules (unitegmic) straight [atropous] (anatropous), (nucellar cap +), (outer and inner integuments ca 5 cells across), micropyle (exo - Idesia)/bistomal/zig-zag, embryo sac elongated, ± protruding into the micropyle, (bisporic, Allium-type), styles separate or style +; fruit also a berry (drupe); seeds arillate (with hairs [of arillate origin], testa also multiplicative but otherwise undistinguished, or exotesta or endotesta [Oncoba] palisade); endosperm + (0 - Salix), (embryo green); n = 9, 10-12, 19.

55[list]/1010: Salix (450: notorious for interspecific hybridisation; connate P modified as nectary, vascular traces to absent members, secondarily insect-pollinated?), Casearia (180), Homalium (180), Xylosma (85). Pantropical, also temperate (but few Australia, not New Zealand) to Arctic (Map: (Africa incomplete) from Sleumer 1954; Meusel et al. 1975; Sleumer 1980; Hultén & Fries 1986). [Photo - Flower, Fruit.]

• Salicaceae can be recognised by their alternate, serrate, stipulate leaves with distinctive "salicoid teeth"; these latter have a small vein proceeding into the tooth, where it expands, the apex of the tooth being a variously coloured spherical gland or stout hair. The flowers often have many spreading stamens and a commonly three-carpellate gynoecium with parietal placentation and either separate styles or a long, persistent style; the fruit is often a capsule. When there are both sepals and petals, the two whorls are in a regular alternating arrangement, and the nectariferous disc or nectar glands and often outside the stamens. A number of taxa have pellucid foliar glands (Abatia, Casearia, Ryania, Zuelania).

Ehrlich and Raven (1964) noted that Atella (Nymphalinae) feeds on Flacourtiaceae and Salicaceae, while some Notodontidae moths (Miller 1992), rusts, e.g. Melampsora spp. on Salix, M. idesiae on Idesia (Holm 1979), etc., show similar host pattern. Boucher et al. (2003) describe Pseudosalix, an Eocene fossil from North America, which is morphologically intermediate between Salix and more morphologically conventional Salicaceae.

Circumscription: Abatieae, Bembicieae, Prockieae (inc. Banareae), Oncobeae (Oncoba only), Homalieae, Saliceae, Samydeae (Casearieae), Scolopieae, and Scyphostegieae (see Chase et al. 2002). Tribal limits may well have to be adjusted, thus Saliceae will probably have to be expanded, Flacourtieae in their current circumscription are polyphlyetic, etc. Trichostephanus (Trichostephaneae) was not assigned to any family (Chase et al. 2002), but in its absence of petals and in having a disc at the base of the calyx it is like Casearia. Oncoba is remarkably like other members of the erstwhile Oncobeae (see now Achariaceae - Lindackerieae), but they differ in chemistry, leaf tooth type, and stamen initiation.

Alford (2003) recognised three families for the New World flacs in addition to Achariaceae, Berberidopsidaceae and Lacistemataceae. His Samydaceae have arillate seeds, punctate or lineate leaves, and flowers with a hypanthium, while the two other families have salicoid leaf teeth, but lack arillate seeds or hypanthial flowers. Salicaeae have dilated stigmas and petals and sometimes sepals absent, while Flacourtiaceae s. str. have attenuate, lobed, or capitate stigmas and petals present or absent.

"Flacourtiaceae as a family is only a fiction; only the tribes are homogeneous" (Sleumer, in Miller 1975). Salicaceae s. str. (cuticle waxes as platelet rosettes, or 0; inflorescences catkins; P much reduced, G [2], collateral, ovules usu. unitegmic; seeds with long placental hairs at base, tegmen not persisting; embryo green) seem quite distinctive. There is a gene duplication in the common ancestor of Salix and Populus, the salicoid duplication, dated to 65-60 million years before present (Tuskan et al. 2006); it will be interesting to know if other Salicaceae s.l. have this duplication. Indeed, it has been observed that Salicaceae s. str. and Flacourtiaceae-Idesieae are very close - they have distinctive leaf teeth, phenolic-type compounds such as salicin are found here only, etc., and as just noted, rusts and caterpillars, perhaps keying in on chemical characters, show similar distributions (see e.g. Meeuse 1975). Elongated embryo sacs occur in both Salicaceae and old Flacourtiaceae (Steyn et al. 2005). Furthermore, it was known that within the old Flacourtiaceae there were two rather different kinds of seed coat (Corner 1976). Chase et al. (2002) have clarified the situation (see also Judd 1997a; Nandi et al. 1998; T. Azuma et al. 2000; Savolainen et al. 2000a), although sampling within tribes still needs to be extended. Some of the old Flacourtiaceae are now in Achariaceae, a few in Lacistemataceae, while Berberidopsideae are in Berberidopsidales (as Berberidopsidaceae) and Aphloia (Aphloiaceae) is placed in Crossosomatales. The remainder of Flacourtiaceae are here. Variation in chemistry, leaf teeth, floral morphology, and seed coat anatomy is largely correlated with this division. Salicoid teeth are quite variable, but all have palisade cells over parenchyma and are well supplied by vascular, especially xylem, tissue (Wilkinson 2007). It should also be noted that Casearia, which may lack salicoid leaf teeth and has apetalous flowers with the disc on the basal-adaxial surface of the calyx, is sister to the rest of Salicaceae, although support for this position is weak (Chase et al. 2002, but cf. D. Soltis et al. 1999, 2000).

Salix and its immediate relatives have ectomycorrhizae. Banara is the only genus reported to have cyanogenic glycosides, but it is well embedded within Salicaceae (Chase et al. 2002). The perforation plates of the tracheary elements are more or less simple and the intervascular pits are small. Xylosma and some Casearia seem to have unilacunar nodes. Leaf traces arise an internode below the leaf they innervate in Hasseltia. Casearia can have phyllanthoid branching, the orthotropic axes having spirally arranged and reduced leaves while the plagiotropic branches are sylleptic and have fully-expanded and 2-ranked leaves. Xylosma, Flacourtia, etc. have groups of large sclereids in the phloem (Zahur 1959). Abatia has opposite leaves with at most very small stipules and marginal glands at the base of the lamina, its valvate perianth members are basally connate and bear many filamentous processes, and it lacks any nectary. There are also taxa with pli-nerved leaf blades and foliar glands, while Populus is dioecious and wind-pollinated... The disc is very variable, and it is often broken up into lobes; it is then sometimes intrastaminal. The embryo sac more or less protrudes into the micropyle in Archevaletaia (Mahshwari 1950). The exotegmen of Dovyalis consists of ribbon-type cells.

Some information is taken from Hegnauer (1973, 1990, also 1966, 1989, as Flacourtiaceae: chemistry), van Heel (1977, 1979: testa anatomy), Miller (1975: wood anatomy), Spencer and Seigler (1985: chemistry), Lemke (1988: general), Gavrilova (1998: pollen), Leskinen and Alström-Rapaport (1999: relationships of Salix and its immediate relatives), and Steyn et al. (2004, 2005: ovule and seed development); see also Judd (1997a) and especially Chase et al. (2002). Bernhard and Endress (1999) discuss androecial initiation. I am grateful to S. Zmarzty for comments.

Synonymy: Bembiciaceae R. C. Keating & Takhtajan, Caseariaceae Doweld, Flacourtiaceae Richard, Homaliaceae R. Brown (G inferior), Poliothyrsidaceae Doweld, Prockiaceae Bertuch, Samydaceae Ventenat, nom. cons., Scyphostegiaceae Hutchinson, nom. cons. (Cork superficial; vessels in radial multiples, with simple (and scalariform) perforations; rays mostly uniseriate; petiole bundle annular and with adaxial mass of xylem and phloem becoming adaxial inverted plate of vascular tissue; stomata paracytic; leaves distichous; plant dioecious, inflorescences terminal, paniculate, long-lived, with large, overlapping, tubular bracts [c.f. Alpinia sect. Myriocrater!], pedicels not articulated; C 3 + 3, connate; staminate flowers: nectariferous lobes opposite A, A 3, connate, extrorse, pollen ?tricolpate; carpellate flowers: G [8-13], ovules basal, micropyle bistomal, funicle well developed, nucellar cap +, stigmas sessile, ray-like, with an opening in the middle; fruit a fleshy capsule with lignified commissural valves; seeds with aril from funicle/outer integument; exotegmic fibers +; endosperm slight, perisperm +, very scanty; n = 9. 1/1: Scyphostegia borneensis. Borneo, not the southern part. [Photo - Flower, Leaf, Inflorescence.] - see Metcalfe 1954 (anatomy); van Heel 1967 (flowers and fruits); Hutchinson 1973 (the diversity of early interpretations of the gynoecium); Chase et al. 2002 (general)]).

LACISTEMATACEAE Martius, nom. cons.   Back to Malpighiales

Trees; plants Al accumulators; chemistry?; vessel elements with scalariform perforations; sieve tubes?; petiole bundle D or deeply C-shaped, also wing bundles +; leaves 2-ranked, (entire); inflorescence a raceme to dense spike; flowers small, P cup-like [1-6], A 1, the thecae ± separated and even stipitate, G [2-3], median member adaxial, 1-2 ± apical ovules/carpel, funicles thick, long, style branches short, ?stigma; fruit a 1(-3)-seeded capsule; testa fleshy or not; endosperm copious, embryo (short), with foliaceous cotyledons; n = ?

2[list]/14. Greater Antilles (Jamaica), Mexico southwards, not in Chile. (Map: from Sleumer 1980).[Photo - Flower, Fruit]

• Lacistemataceae are woody, evergreen plants that have stipulate and usually serrate leaves. The small flowers are borne in rather dense racemes to spikes and have an inconspicuous perianth and a single anther whose two thecae are often well separated and even stipitate. The dehiscent fruit is small, with one, sometimes to three, apical seeds.

Lacistemataceae do not cluster with the rest of Salicaceae and Kiggelariaceae (Savolainen et al. 2000a; Chase et al. 2002), although they are probably in this area (Chase et al. 2002; see also D. Soltis et al. 1999, 2000). Davis et al. (2005a) place them as sister to Salicaceae s.l. (61% bootstrap, 100% posterior probability), as do Korotkova et al. (2007: slightly higher jacknife); as might be expected, they lack salicoid teeth.

Is there an aril in Lacistemataceae? Sleumer (1980) records one for Lacistema, but a fleshy seed coat for Lozania; an aril is obvious in neither. In the latter genus there appear to be long "hairs" inside the fruit which perhaps support the dangling seed; these hairs are thick-walled but unlignified cells that may be derived from the funicle.

See Sleumer (1980: as Flacourtiaceae - Lacistemeae) for a monograph. Lozania - Riviere 270 (anatomy), Gentry et al. 22231 (fruit); Lacistema - Aymard & Delgado 6882 (fruit), Rimachi Y. 11201 (anatomy - stomata tending to anisocytic).

Turneraceae + Malesherbiaceae + Passifloraceae [= Passifloraceae s.l.]: cyclopentenoid cyanogenic glycosides and/or cyclopentenyl fatty acids +; (plant with unpleasant smell); (foliar glands +), colleters +; K + C together forming a tube, (corona or scales on C), A 5 [opposite K], styles +; aril +, endotestal cells large, exotegmen palisade, endotegmen persistent; endosperm persistent, oily.

The cyclopentenoid glycosides may be sequestered by caterpillars feeding on plants of these families and perhaps even used as nitrogen sources; Achariaceae also have this combination of features.

Turneraceae are weakly associated with Malesherbiaceae in Chase et al. (2002), the two being strongly associated with Passifloraceae. Turneraceae show biparental or paternal transmission of plastids, as may Passifloraceae (Shore et al. 1994). Including Turneraceae and Malesherbiaceae in Passifloraceae s.l. is an optional arrangement in A.P.G. II. However, basic clade limits within this group may need re-evaluating, since preliminary data suggest that a paraphyletic Passifloraceae may include Turneraceae and Malesherbiaceae (A.P.G. II 2003), while Korotkova et al. (2007: three taxa from the three families) found Turnera and Passiflora to be sister taxa with 98% jacknife support.

For a discussion on aril development in this group, see Kloos and Bouman (1980); although it is often described as funicular, they incline to call it raphal.

Turneraceae + Malesherbiaceae: leaves spiral; exotestal cells arranged in lines; x = 7.

TURNERACEAE Candolle, nom. cons.   Back to Malpighiales

Herbaceous or woody; plant with unpleasant smell; ellagic acid 0; cortical vascular bundles [= leaf traces] common; vessel elements with simple (and scalariform) perforation plates; stomata various; leaves conduplicate, glands often at base of lamina, stipules 0 (+ - e.g. Erblichia); bracteoles often large; (flowers heterostylous), C contorted, deliquescent, (glands or corona at mouth of tube), nectary near base of tube (on sepals; filaments), microsporogenesis simultaneous, G [2, 3] (half inferior), micropyle zig-zag, hypostase +, stigmas concave, often ± penicillate; K + C tube deciduous; aril +; seeds often regularly rugose; n also = 5 (13).

10[list]/205: Turnera (122), Piriquetia (44). Tropical to warm temperate America and Africa (inc. Madagascar and Rodriguez I.) (Map: from Wickens 1976; Heywood 2007). [Photo - Flower]

• Turneraceae are herbs to woody plants that may be recognised by their often very hairy, serrate, exstipulate leaves with strong secondary venation. The flowers, with their calyx and corolla together forming a tube, contorted and deliquescent yellow or sometimes red corolla, parietal placentation, and often bifid, fringed stigmas, are distinctive. The loculicidally capsular fruits have arillate seeds with a minutely and often regularly reticulate testa.

Turneraceae are the hosts of caterpillars of several genera of Nymphalidae, alternate hosts include Salicaceae, Passifloraceae, and Violaceae (Arbo 2006 and references). It has been suggested that there is floral mimicry between Turnera and Malvaceae in Argentina (Benitez-Vieyra et al. 2007). Species of both Turnera and Piriqueta have epiphyllous flowers. Heterostyly is common in these and some other genera of Turneraceae.

Some information is taken from Vijayaraghavan and Kaur (1967: embryology), Hegnauer (chemistry), González and Arbo (2005: anatomy), and Arbo (2006: general account).

Synonymy: Piriquetaceae Martynov

MALESHERBIACEAE D. Don, nom. cons.   Back to Malpighiales

Herbaceous or subwoody; hairs conspicuous, multiseriate, often glandular; plant with unpleasant smell; tannins?; (cork cortical); vessel elements usu. with simple perforations; nodes also 1:1; leaves often deeply lobed, (margins entire; stipules foliaceous or 0), colleters?; K + C tube long, K valvate, C valvate, (denticulate corona at the mouth of the tube), androgynophore +, G [3, 4], micropyle endostomal, large protrusion from chalazal region, styles slender, stigmas capitate-clavate, ?type; K + C tube persistent; seeds pitted, aril 0; endosperm type?

1[list]/24. Andean South America from Peru S., esp. N. Chile (Map: see Gengler-Novak 2002). [Photo - Habit]

• Malesherbiaceae are more or less herbaceous, foetid and often densely glandular-hairy plants with distinctive flowers. The calyx and corolla together form a long tube; the sepals and petals are both valvate. There is an androgynophore, the floral tube is persistent in fruit, and the styles are often subapical on the capsule.

For relationships within the family, see Gengler-Novak (2002, 2003).

Cronquist (1981) suggests that stipules are lacking. Malesherbiaceae are very close to Turneraceae, but their seeds lack arils, the corolla aestivation differs, etc.; they are also close to Passifloraceae (Chase et al. 2002). The styles are shown as being commissural by Schnizlein (1843-1870: fam. 198).

General information is taken from Ricardo S. (1967, he suggested that the micropyle is endostomal) and Kubitzki (2006b); for chemistry, see Hegnauer (1969, 1990).

PASSIFLORACEAE Kunth, nom. cons.   Back to Malpighiales

Woody or herbaceous climbers with simple branch tendrils; flavonols +, ellagic acid +/0, tannins 0; anomalous secondary thickening quite common; vessel elements with simple perforation plates; wood often fluorescing; supernumerary buds +; leaves spiral, (compound), conduplicate, (margins entire), 2ndary veins often palmate, glands common on petiole or lamina; (dioecious), inflorescence cymose; flowers (3-)5-merous, corona of (1-)2-several rows of filaments or membranes (0), nectary ± on K/C tube, (A basally connate), gynophore or androgynophore frequent, anthers versatile, pollen to 12-colporate, G [(2-)3(-7)], funicle often long, nucellar apex pointed, micropyle bistomal or zig-zag, stigmas capitate or divided [Adenia], papillae multicellular; fruit usu. a berry; seeds often flattened, sculpted, bony, hairy or not, ruminate, aril +, testa multiplicative, sarcoexotestal or not, endotesta crystalliferous, lignified or not; n = 6 (7) 9(-12).

16[list]/705 (10/675): Passiflora (525, not Africa, inc. Hollrungia and Tetrapathea - Krosnick & Freudenstein 2006), Adenia (100: see Hearn 2006 for a phylogeny, great variation in life form). Tropics to warm temperate, especially Africa and America (Map: from van Balgooy 1975; George 1982). [Photo - Collection]

Paropsieae

Trees or shrubs; vessel elements in multiples, with scalariform perforation plates; leaves spiral, reduced [orthotropic axes], or 2-ranked [plagiotropic axes], 2ndry venation pinnate, (stipules 0); inflorescence racemose; (A -20), androgynophore +/0, (pollen 6-porate), nectary 0 (annular), G [(2-)3-6], (style + [Barteria]); seeds not sculpted; n = ?

6/ca 27: Paropsia 12. Tropical, esp. (west) Africa.

• Many Passifloraceae are tendrillar vines or lianes; the leaves are alternate, stipulate, often with palmate venation, and often with glands on the lamina or petiole. The tendril is often oblique to the axis; a flower may be lateral to it, or paired flowers are borne on it or at its base. The axillary buds are superposed, and the lower bud produces the inflorescence/tendril. Floral features are, however, more important for the recognition of the family. The calyx and corolla are often similar and fused to each other at the base, and there is usually a complex corona. There is a gynophore or androgynophore, often five stamens with large and versatile anthers, and a three-carpellate gynoecium with parietal placentation. The fruit is a berry, and the often flattened seeds are arillate.

Passiflora and its relatives are known for their association with Heliconius butterflies, and the former show great variation in leaf morphology, foliar glands (some of these are involved in egg mimicry - Vanderplank 2007 for references), etc. Heliconius itself is also closely associated with Anguria (Cucurbitaceae), which it pollinates, also obtaining nutrients from the pollen (e.g. Gilbert 1972, 1975; Spencer 1988). The larvae of some Acraeinae and also of brightly-coloured Notodontidae-Dioptininae moths are also often found on Passiflora (Miller 1992; Silva-Brandão et al. 2008), and at least the former are also found on Barteria. Details of the association between the African ant-plant Bartera fistulosa and the ant Tetraponera aethiops are given by Dejean et al. (2008).

Flacourtiaceae - Paropsieae (Barteria, Paropsia, etc.) are to be included here (Chase et al. 2002); I have characterised them separately above. Anatomically the two are rather similar, indeed, the major variation in the family seems to be associated with habit - lianes versus tress (Ayensu & Stern 1964). Passiflora and immediate relatives have stem collenchyma, cymose inflorescences, and branches developing from an accessory (superposed) bud, as is common in taxa with axillary tendrils with non-basal prophyllar buds, which in this case may be flowers. Adenia may be rather different from other Passifloraceae, perhaps more like Malesherbiaceae and Turneraceae, e.g. in moderately developed corona, tricolporate pollen (e.g. see Feuillet & MacDougal 2006). Indeed, Adenia has a nectary often made up of separate glands, a hollow style, and its stigma lacks multicellular papillae (Bernhard 1999a, c), in addition, it may be dioecious, it lacks an androgynophore, the stamens are sometimes connate, and there is a gynophore; some species have a true hypanthium (de Wilde 1971b)!

In species of Passiflora with strongly bilobed leaves, ptyxis may be modified conduplicate: the blade makes a V with an inverted V at the end of each arm. The tendril is an axillary shoot and flowers can arise from prophyllar buds. Sazima and Sazima (1978) note that the bat-pollinated flowers of Passiflora mucronata become zygomorphic as the stamens move after the flowers opens; Endress and Matthews (2006a) give this as an example of monosymmetry in the family. For floral anatomy of Passiflora, see Puri (1947), for a general account of the genus, see Ulmer and MacDougal (2004), for a formal infrageneric classification, see Feuillet and Macdougal (2004), for floral development, see Krosnick et al. (2006), and for the phylogeny of Passiflora, see Yockteng and Nadot (2004), Krosnick and Freudenstein (2005: also morphology). Hansen et al. (2006) discuss chromosome number evolution, n = 12 may be the basal number; see also de Melo and Guerra (2003).

Do the sieve tubes have non-dispersive protein bodies?

Carpel orientation is taken from Le Maout and Decaisne (1868) and Schnizlein (1843-1870: fam. 197); for pollen, see Presting (1965) and Spirlet (1965), for chemistry, see Hegnauer (1969, 1990), for embryology, etc., see Singh (1970), for branching, see de Wilde (1971a), for general information, de Wilde (1974), for anatomy, see Harms (1893), for stipules, see Dahlgren and van Wyk (1988), and for a general account, see Feuillet and MacDougal (2006). I am grateful to J. M. MacDougal for information.

Synonymy: Modeccaceae Horaninow, Paropsiaceae Dumortier, Smeathmanniaceae Perleb

GOUPIACEAE Miers   Back to Malpighiales

Evergreen trees; plants Al-accumulators, otherwise chemistry unknown; vessel elements with scalariform perforations; nodes ?; petiole with ± annular bundle(s), plus inverted medullary bundles; branched sclereids +; cuticle waxes 0; leaves 2-ranked, tooth ?type, 2ndary veins actinodromous, 3ary veins scalariform; inflorescences axillary, umbellate, pedicel articulation?; C induplicate-valvate, long, apical part inflexed, disc annular, connective shortly prolonged, with long hairs, pollen with endexinal folds, G opposite C, placentation axile, several basal ovules/carpel, styles separate, short, on outer shoulders of carpels, stigmas subulate, type?; fruit a drupe; testa and tegmen ca 6 cells thick, testa with one layer [mesotestal] of sclereids, exotegmen poorly developed; endosperm copious; n = ?

1/2. N.E. South America.

• Goupiaceae may be recognised by their two-ranked stipulate leaves with ascending secondary veins and scalariform tertiaries. The inflorescence is umbellate and the petals are relatively long and inflexed at the apex. The gynoecium has basal-axile placentation and widely separated styles; the fruit is a single-seeded drupe.

The family is poorly known. Cronquist (1981) included Goupiaceae in Celastraceae, Takhtajan (1997) in Celastrales, A.-L. de Jussieu and others have placed it in Rhamnaceae. It is often suggested that only seedlings have dentate leaves, those of the adult being entire, but leaves of flowering specimens are in fact frequently toothed.

Information on pollen is taken from Lobreau-Callen (1980), and on seed, from Takhtajan (2000).

Lophopyxidaceae + Putranjivaceae: stomata paracytic; 2 apical epitropous ovules/carpel; fruit 1-seeded.

This clade may have separated in the Cretaceous-Albian 111-100 million years before present, Lophopyxidaceae and Putranjivaceae themselves diverging at end Coniacian or thereabouts ca 85 million years before present (Davis et al. 2005a).

Davis et al. (2005a) find a strong association between these two families; they may in turn be associated with the group of families with parietal placentation.

LOPHOPYXIDACEAE H. Pfeiffer   Back to Malpighiales

Liane with leaf tendrils; chemistry?; branches with lateral bud at base; secondary thickening with included phloem; vessel elements with simple perforations; phloem stratified; nodes ?; petiole bundles arcuate; leaves spiral; plant monoecious, inflorescence branched, flowers in clusters, sessile, small, K connate basally, valvate, C very small, staminate flowers: A = and opposite K, cordate glands adnate to C; carpellate flowers: glands forming a lobed disc, G [(4) 5], opposite petals, small funicular obturator +, style short, branched, stigmas subulate; fruit a 5-winged samara; seed coat?; endosperm ?development, +; n = ?

1/1: Lophopyxis maingayi. Malesia to the Solomon and Caroline Islands (Map: from Sleumer 1971b).

• Lophopyxidaceae are tendrillate lianas with quite strongly serrate leaves that have pinnate venation; the flowers are small and are borne in tight clusters along the branched inflorescences. The one-seeded fruit has five wings.

Included in Celastraceae by Cronquist (1981) and Hutchinson (1973), in Celastrales by Takhtajan (1997), but placed close to Pandaceae (represented by Microdesmis) by Savolainen et al. (2000a; see also Chase et al. 2002). Sleumer (1971b) described the tendrils as being leaves and also bud-bearing branches; the ultimate spirally-recurved portion seems to be foliar.

PUTRANJIVACEAE Endlicher   Back to Malpighiales

Evergreen trees; cucurbitacins [triterpenes], glucosinolates, biflavonoyls +; cork?; vessel elements with scalariform perforation plates; petiole bundles elliptic; hairs unicellular; leaves 2-ranked, (veins running into opaque deciduous teeth, or spines); plant dioecious, inflorescence fasciculate; P 4-5(-7), staminate flowers: A (2-)3-20(-many; extrorse), disc + or 0; carpellate flowers: G 1[-4(-9)], ovules epitropous, nucellus ca 2 cells thick, disintegrating early, micropyle (exo/)endostomal, outer integument 3-8 cells and inner 6-14 cells across, archesporium 2-3-celled, endothelium +, placental obturator +, style branches short or 0, stigmas flap-like, ?type; fruit a drupe; testa vascularised, exomesotesta sclereidal, tegmen 6-24 or more cells thick, exotegmen cells cuboidal; endosperm copious; n = (19) 20 (21).

3[list]/210: Drypetes (200). Tropical, esp. Africa and Malesia (Map: from FloraBase 2005; Andrew Ford, pers. comm). [Photo - Flower, Fruit]

• Putranjivaceae are trees to shrubs that may be recognised by their 2-ranked often rather coriaceous leaves that are asymmetrical at the base and dry a greyish color; they frequently taste peppery or like an inferior radish when fresh, although the taste may take a little time to develop (mustard oils!). The fasciculate flowers are frequently rather small and the single-seeded drupe is often crowned by the persistent, flap-like stigmas.

Perhaps not surprisingly, caterpillars of Pierid butterflies have quite often (23/2690 records) been recorded from this group (see also Brassicales and Fabaceae) - nothing so far known from Lophopyxidaceae!

Putranjivaceae have usually been included in Euphorbiaceae (as by Webster 1994, in Phyllanthoideae), but can be distinguished i.a. by their chemistry, embryology, and fruit. They are also certainly not to be placed with the rest of the glucosinolate families in Brassicales (e.g. Rodman et al. 1997, 1998).

For chemistry, see Hegnauer (1966, 1989, as Euphorbiaceae), for embryology and seed anatomy, see Singh (1970), Stuppy (1996), and Tokuoka and Tobe (1999, 2001 - Lingelsheimia included, but tegmen 3-4 cells thick and testa vascularised, to be placed in Phyllanthaceae - see Kathriarachchi et al. 2005), for wood anatomy, Hayden and Brandt (1984 - it is like that of Aporusa, etc. [= Phyllanthaceae]). For a checklist and bibliography, see Govaerts et al. (2000).

Erythroxylaceae + Rhizophoraceae: tropane [hygroline] and pyrrolidine alkaloids, non-hydrolysable tannins +; sieve tube plastids with protein crystalloids; mucilage cells common; stomata paracytic; leaves opposite, involute, colleters +, stipules enclosing the terminal bud; inflorescence cymose; K valvate, C conduplicate, enclosing A, median G adaxial, 2 pendulous epitropous ovules/carpel, inner integument ca 6 cells thick, endothelium +; fruit a septicidal capsule, K persistent; seeds arillate, exotestal cells enlarged, thick-walled, ± tanniniferous; endosperm +, starchy, embryo green.

This clade may have diverged in the Cretaceous-Aptian around 114 million years before present ([119.3-]113.8[-110.2]/[105.7-]101.6[-102.1] million years before present: Davis et al. 2005a).

There is weak support for an association of [Caryocaraceae [Linaceae + Irvingiaceae}} with this clade (Soltis et al. 2007a), and there are a number of features in common, such as a basally connate androecium, epitropous ovules with an endothelium, etc. (Matthews & Endress 2007). Ctenolophonaceae, etc., may also be associated, but its floral similarities are not so great.

Including the two families in Rhizophoraceae s.l. is optional in A.P.G. II, but the two are so well known and for the most part distinctive that keeping them separate is preferable. Rhizophoraceae used to be placed in Myrtales (Cronquist 1981) or Myrtanae (Takhtajan 1997), largely because of their vestured pits and inferior ovary, but they are well supported as sister to Erythroxylaceae (e.g. Setoguchi et al. 1999; Schwarzbach & Ricklefs 2000; Chase et al. 2002; Korotkova et al. 2007). Although an unexpected family pair, when comparing Aneulophus (Erythroxylaceae) with non-mangrove Rhizophoraceae, the differences are then less obvious.

For floral development, see Matthews and Endress (2007).

ERYTHROXYLACEAE Kunth, nom. cons.   Back to Malpighiales

Smallish trees and shrubs; ellagic acid 0; vessel elements with simple perforations; wood commonly with SiO₂ grains; nodes with lateral bundles originating well before the central, forming cortical bundles; cristarque cells in bundle sheath and ground tissue (0); sclereids +; petiole bundle arcuate to annular with medullary and adaxial bundles; branching from previous flush; buds perulate; leaves also usu. 2-ranked (spiral); inflorescence often fasciculate; pedicel articulated (or not - Aneulophus?), heterostyly common, K connate basally, C with fringed bilobed ligule (0), A 10, obdiplostemonous, extrorse, ± connate (at base only - Aneulophus), connective often thickened, pollen trinucleate, nectary 0, G [(2-)3(-4)], (adaxial only fertile), also 1 ovule/carpel, micropyle endostomal, style (short), (hollow), branches separate, stigmas ± capitate; fruit often a 1-seeded drupe, A also persistent; (endosperm 0); n = 12.

4[list]/240: Erythroxylum (230). Pantropical, esp. American (Map: from van Steenis and van Balgooy 1966; Heywood 1978). [Photo - Flower, Fruit.]

• Erythroxylaceae are shrubs to trees that can be recognised by their entire leaves, stipules that are at least in part interpetiolar, and usually fasciculate inflorescence with rather small flowers; both calyx and filaments persist at the base of the fruit. Erythroxylum itself has intrapetiolar stipules, there are often longitudinal markings on the leaves, the filaments are often of two different lengths and are basally connate, and the fruit is a drupe.

Cocaine is sequestered by the larvae of Eloria noyesii, a lymanitrid moth.

Aneulophus has a thick testa, thin tegmen, aril, opposite leaves, colleters, inter/intrapetiolar stipules, and a septicidal capsule; from petal length, it appears that the flowers are monosymmetric.

The leaves of Erythroxylum coca were described as being revolute by Cullen (1978); they are involute (e.g. Peyritsch 1878; Keller 1996). The nodes were described as being unilacunar by Sinnott (1914). Erythroxylum sometimes has milky exudate. Are the lamina teeth theoid??

For chemistry, see Hegnauer (1966, 1989).

Synonymy: Nectaropetalaceae (H. Winkler) Exell and Mendonça

RHIZOPHORACEAE Persoon, nom. cons.   Back to Malpighiales

Trees; ellagic acid +; vessel elements with simple and/or scalariform perforations; true tracheids +; pits vestured; nodes also multilacunar, often with split-laterals; subepidermal laticifers in flower; branching from current flush; leaves (supervolute - mangroves), (margins entire); inflorescence axis often evident; K (3-)4-5(-16), C small, often hairy, variously lobed, fringed, or with filiform appendages, aristate, enveloping 1-small groups of A, A 9-many, (connate), disc on ovary or hypanthium (0), G [2-many], opposite K, when 2, lateral, micropyle zig-zag or endostomal, inner integument 5-20 and outer integument 2-5 cells across, several megasporocytes, style + (branched - Gynotroches), stigma ± punctate, capitate or lobed, ?type; (endotesta crystalliferous); (embryo short), cotyledons large; n = (13), 14, 16, 18, 21; germination epigeal, cotyledonary node unilacunar.

16[list]/149. Three groups below. Pantropical (Map: from Ding Hou 1958; van Steenis 1963; George 1984; Tomlinson 1986; Juncosa & Tomlinson 1988a). [Photo - Flower. Flower, Fruit.]

1. Macariseae

Crystals solitary; stipules valvate; (seeds winged at micropylar end).

7/94: Cassipourea (62), Dactylopetalum (15). Tropical America and Africa, also peninsula India and Sri Lanka.

Gynotrocheae + Rhizophoreae: stilt roots present; rootlets without root hairs; leaves bijugate, stipules imbricate; hypanthium +, ovary ± inferior.

2. Paradrypetes

Raphides +; long, zig-zag intersecondary veins; plant dioecious; flowers small, P 3-4, pollen grains spiny, nectary 0, placental obturator +, style 0; fruit a drupe; seed coat vascularised; abundant starchy endosperm +.

1/2. Brasil.

Gynotrocheae + Rhizophoreae: stilt roots present; rootlets without root hairs; leaves bijugate, stipules imbricate; hypanthium +, ovary ± inferior.

3. Gynotrocheae

(Stilt roots 0); (to 8 tenuinucellate ovules/carpel, archesporial cells single); fruit a berry; meso- and endotegmen persist.

4/30: Crossostylis (10). Indo-Malesia, Madagascar.

4. Rhizophoreae

Stomata cyclocytic; abaxial hypodermis +; sclerenchymatous sheath of midrib at most weakly developed; leaves entire; endothelium 0; fruit indehiscent, 1-seeded; seed coat undifferentiated, vascularized, tegmen not persisting; (endosperm overflows from embryo sac); seeds germinating on tree; cotyledonary node tri- or multilacunar.

4/17: Rhizophora (?9). Pantropical, but centred on the eastern Indian Ocean; see the blue area in the map above.

• Rhizophoraceae are shrubs to trees that may be recognised by their opposite often serrate leaves and large interpetiolar stipules. The flowers are quite often other than 5-merous, the calyx is thick and valvate, the petals are often lobed or fringed, each enclosing one or small groups of stamens, and the ovary is often more or less inferior. Gynotrocheae and Rhizophoreae have aerial prop roots.

Mangrove taxa, derived within the family (e.g. Schwarzbach & Ricklefs 2000), are most diverse in Southeast Asia-Malesia. Pollen is deposited on to the hairy petals, so there may be secondary pollen presentation, but pollination is basically explosive, the stamens being held in groups by the petals until the flower is tripped by the pollinator. The pollen grains are very small, and in Rhizophora in particular pollination may be by wind (Juncosa & Tomlinson 1988b). (These petals often have an arista or other appendages, and are shaped like a tiny bivalve mollusc - Endress & Matthews 2006b.) Mangrove seeds have little endosperm and are viviparous (aquatic/marine plants quite commonly have large embryos), and in all genera except Bruguiera the endosperm overflows from the seed, pushing open the micropyle as it does so. After the seed falls from the tree it may float in the water, the hypocotyl straightening and establishment of the seeding being by the development of lateral roots (Juncosa & Tomlinson 1988b). Depending on the genus, there are either stilt roots or pneumatophores, and axillary buds soon die so the plants cannot regenerate when cut (or if the twigs are killed by frost), etc. (see Tomlinson 1986 for much useful information). Rhizophora is known from the Caribbean more or less continuously since the late Eocene, although the common ancestor of the existing populations there may have arrived in the New World some 40 million years later, ca 11 million years before present (Graham 2006).

Schwarzbach and Ricklefs (2000) found strong phylogenetic structure in the family, and suugested the recognition of three tribes. Crossostylis, with dehiscent fruits and arillate seeds, is embedded in Gynotrocheae, which otherwise have fleshy, indehiscent fruits and seeds without arils. Have fleshy indehiscent fruits evolved in parallel within Gynotrocheae, or is the arillate seed, etc., of Gynotroches a reversal? Molecular data also place Paradrypetes (ex Euphorbiaceae) here (e.g. Davis et al. 2005a), for data on that genus, see Levin (1986, 1992) and Radcliffe Smith (2001 - as Euphorbiaceae). It has a rather unexpected combination of characters.

Growth in Rhizophoraceae may be continuous, although it appears the growth patterns in Macarisieae are unknown. The lead teeth are theoid. Rhizophora has transverse carpels (Eichler 1876). Cork initation in the root may be superficial for at least some taxa, perhaps just those with stilt roots (see von Guttenberg 1968 for Carallia).

See also Juncosa and Tomlinson (1988: general), Tobe and Raven (1988: seed coat anatomy), Endress and Matthews (2006b: petal morphology) and Baranova and Jeffrey (2006: leaf anatomy).

Synonymy: Cassipoureaceae J. Agardh, Macarisiaceae J. Agardh, Mangiaceae Rafinesque

LINACEAE Perleb, nom. cons.   Back to Malpighiales

Cork?; vessel elements with simple or scalariform perforations; true tracheids +; petiole bundle(s) arcuate; cristarque cells + [rare in Linoideae]; epidermal wax crystals as parallel platelets; branching from previous innovation; leaves usu. involute, tooth ?type, petiole short; pedicels articulated; K quincuncial, C contorted, caducous, A basally connate, pollen starchy, G [2-5], opposite C, or median member adaxial, (1) 2 apical ovules/carpel, placental obturator +, style more or less divided, stigmas capitate; fruit various, K persistent; tegmen strongly multiplicative; endosperm variable, chalazal haustorium +, (embryo slightly curved).

10-12[list]/300 - two groups below. World-wide, trees and lianes tropical. (Map: from van Hooren & Nooteboom 1984a; Hultén & Fries 1986; Jardim 1999; Diderichsen & Richards 2003: Hugonioideae tropical, area outlined in black, Linoideae more temperate, tropical areas outlined in blue. China, etc., incomplete.)

1. Linoideae

Herbs (shrubs); ellagic acid 0; vessels with simple perforation plates; rays uniseriate; nodes 1:1 [Linum]; cuticle waxes as parallel platelets; leaves opposite or spiral, margins entire or toothed (stipules 0); K ± equal, C clawed, nectary outside filaments or at base of C, A 5, opposite K, alternating with staminodes, pollen trinucleate, ?often starchy, 3-pantocolpate or pantocolporate, or inaperturate, G loculi usu. divided, (archesporium multisporangiate), ovules tenuinucellate, inner integument 9-12 cells across, endothelium +, stigma wet or dry; fruit a septicidal capsule, (2-seeded mericarps splitting along false septae, units opening adaxially); seeds often mucilaginous, exotesta with outer walls massively thickened, cross cells beneath exotegmen, endosperm scanty, (helobial), embryo green [Linum]; n = 6, (8)9(1-18, etc.).

6/240: Linum (180). Worldwide, but esp. N. temperate and subtropical. [Photo - Flower]

2. Hugonioideae

Woody, often lianes with branch grapnels; ellagic acid?; vessels with scalariform perforation plates; sclereids +; stomata accessory cells usu. lignified, lobed beneath the guard cells; branching from previous flush; leaves spiral or 2-ranked, margins toothed, (stipules pectinate); K often unequal, C at most slightly clawed, often yellow, disc at base of filaments, A 10, of two lengths, micropyle endostomal [Roucheria]; fruit a drupe or with mericarps; seed with at most slight arillode, mesotesta with sclerotic cells, endotesta lignified, exotegmen barely lignified or tegmen obliterated; endosperm copious to scanty; cotyledons large; n = 6, 12, 13.

4-6/61. Pantropical. [Photo - Flower]

Only half of the ovules may develop and produce seeds. Note that septicidal dehiscence, presumably liberating pyrenes, may occur in this subfamily (Spichiger et al. 2002).

Synonymy: Hugoniaceae Arnott

• The leaf may have longitudinal lines on the abaxial surface, and the petals are often yellow. I find Linaceae difficult to identify...

Linaceae may have diverged in the Cretaceous-Albian 111-100 million years before present (Davis et al. 2005a).

Ellagic acid is not reported from Linoideae, but members of this subfamily are largely herbaceous. Flat ptyxis is reported from Linum narbonense by Cullen (1978), other taxa may be conduplicate. Tirpitzia bilocularis has a corolla tube over 2 cm long. Anisadenia, with its spicate inflorescence, stamens opposite the petals, and 2-carpellate gynoecium, each carpel having a single seed, is very distinctive (Brummitt 2007).

Linaceae are weakly associated with Picrodendraceae in Chase et al (2002a), and with Irvingiaceae in Tokuoka and Tobe (2006).

For information on the family and its possible segregates see Hegnauer (1966, 1989: chemistry), Narayana (1970: embryology, etc.), Robertson (1971: Linoideae), van Hooren and Nooteboom (1984, 1988a, b: general), van Welzen and Baas (1984: anatomy), and Jardim (1999: New World Hugonioideae).

IRVINGIACEAE Exell & Mendonça   Back to Malpighiales

Trees; ellagic acid +; elements with simple perforations; nodes ?; petiole bundle annular; cristarque cells widespread; (sclereids +); mucilage cells in epidermis and ducts elsewhere in leaf; stomata paracytic; branching from previous flush; leaves 2-ranked, revolute, margins entire, 2ndary veins strong, rather close and subparallel, tertiary veins also ± parallel and at right angles to the secondary veins, stipules large, intrapetiolar and encircling terminal bud; inflorescences axillary or terminal panicles; flowers small, pedicels articulated, A (9) 10, filaments folded in bud, disc massive, G [2, 5], G median (when 2) or opposite K, 1 pendulous ovule/carpel, placental obturator +, style +, stigma punctiform, ?type; fruit a drupe [stones 1 or 5 1-seeded, or 5-seeded] or samara; testa thick, much sclerotised; endosperm slight to copious; cotyledons large, cordate; n = ?

3 (Klainedoxa, Desbordesia, Irvingia)[list]/10. Africa; South East Asia to W. Malesia (Map: from Harris 1996). [Photo - Fruit]

• In Irvingiaceae, the leaves, with their longitudinal markings, their large, deciduous, intrapetiolar stipules that enclose the prominent, pointed terminal bud, and their rather closely parallel secondary venation, are distinctive. The axillary buds are also prominent, and prophylls are basal on the shoots. The pedicels are articulated at the junction with the axis; the thin calyx persists on the fruit and is conspicuously reflexed, and the disc is very obvious. The drupes often have radial fibers.

Irvingiaceae may have diverged in the Cretaceous-Albian some 111-100 million years before present (Davis et al. 2005a).

Irvingiaceae are included in Rutales by Takhtajan (1997). Irvingia is sister to Erythroxylum in a tree presented by Fernando et al. (1995); the stipules of Irvingiaceae and Erthroxylaceae (and Linaceae-Ixonanthoideae) are similar (Weberling et al. 1980), but they are weakly associated with Putranjivaceae in Chase et al (2002a) and with Linaceae in Davis et al. (2005a).

Keller (1996) suggests that the leaves are involute in bud.

Information is taken from Jadin (1901: anatomy), Noteboom (1967: chemistry), Harris (1996; monograph), Link (1992c), and Boesewinkel (1994: see tegmen!); details of floral orientation are taken from Eckert (1966).

IXONANTHACEAE Miquel, nom. cons.   Back to Malpighiales

Woody; ellagic acid +; vessel elements with simple perforations; mucilage cells +; cuticle waxes as variously arranged platelets; stomata paracytic; petiole bundle arcuate; branching from previous flush; leaves spiral, involute, (margins entire), (stipules cauline); inflorescences axillary, corymbose; K usu. basally connate, C contorted or imbricate, A 5[opposite K]-20, folded in bud, pollen with supratectal spines, free disc nectaries +, G [(2-)5], (carpels subdivided), (1- Allantospermum)-2 apical ovules/carpel, placental obturator, endothelium, hypostase +, micropyle bistomal, style undivided, slender, stigma capitate or discoid; fruit a septicidal (and loculicidal) capsule opening adaxially as well, columella persistent or not, K and C persistent; seeds basally winged, or aril arising between the hilum and micropyle; endotegmen with sinuous anticlinal walls; cotyledons large; n = ?

4-5[list]/21. Pantropical (Map: from Aubréville 1974; Kool 1988).

Ixonanthaceae may have diverged in the Cretaceous-Albian 111-100 million years before present (Davis et al. 2005a).

The stamens opposite the petals in Ixonanthes are paired, but arise from a single trace (Narayana & Rao 1966). Narayana (1970) depicts a tegmen ca 4 cells thick, the innermost layer of which is an endothelium.

Robson and Airy Shaw (1962) drew attention to the "spiral convolution of the filaments and style" of Cyrillopsis which, they thought, were points of similarity between this genus and Irvingiaceae sensu stricto. However, Allantospermum and some species of Ochthocosmus also have flowers very similar to those of Cyrillopsis, with the thin calyx reflexed after anthesis (Phyllocosmus, Ixonanthes), while other species of Ochthocosmus have persistent, erect, almost scarious-looking sepals, as is common in Linaceae. Takhtajan (1997) included Allantospermum in Irvingiaceae - it has flowers with two carpels and seeds with copious endosperm, and the inflorescences of some Ixonanthaceae are indeed very like those of Irvingiaceae... On the other hand, Bove (1997) suggested that Ixonanthaceae and Humiriaceae were sister taxa, both having ellagic acid, a "free" disc encircling the ovary, and an entire stigma. In the context of Linales (Linaceae and their immediate relatives), Ixonanthaceae were distinct in their free stamens, semi-inferior ovaries and pollen grains with supratectal spines. Davis et al. (2005a) find a week association between Ixonanthaceae and the Clusiaceae group, and Tokuoka and Tobe (2006) a weak association betwee Ixonanthaceae and Ochnaceae.

See also Nooteboom (1967: chemistry), Forman (1965: general), Narayana (1970: embryology, etc.), and Kool (1980: revision of Ixonanthes, 1988: general) for information.

CTENOLOPHONACEAE Exell & Mendonça   Back to Malpighiales

Woody; ellagic acid?; vessel elements with scalariform perforations; oxalate crystals +; cristarque cells +; cuticle waxes 0; stomata anomo- or anisocytic; hairs tufted; petiole bundle arcuate; leaves opposite, entire, stipules interpetiolar; inflorescence terminal, ?thyrsoid, pedicels articulated; C contorted, caducous, disc, A 10, adnate to base of disc, of two lengths, anthers with broad connective, pollen 3-9 zonocolporate, G [2], 2 apical ovules/carpel, with placental obturator, style +, styles short, stigmas capitate; fruit a ? capsule, K persistent, swollen; seed single, persisting on columella; arillode ± hairy, exotestal cells palisade, the outer wall thickened; endosperm copious, cotyledons very large, folded; n = ?.

1[list]/3. W. Africa, Malesia (Map: from van Hooren & Nooteboom 1988b).

• Ctenolophonaceae are trees that may be recognised by their opposite, entire leaves with interpetiolar stipules and their terminal inflorescences with rather elongated flower buds. The rounded sepals are persistent and accrescent, and the grey-drying and closely-ribbed fruit is distinctive. There is a single seed half enveloped by the hairy arillode; it remains attached to the slender central columella.

Ctenolophonaceae may have diverged in the Cretaceous-Albian 111-100 million years before present ([109.6-]101.8[-96.6]/[97.1-]91.0[-88.1] million years before present: Davis et al. 2005a). Fossil pollen is known from South America and India, and the first records are from Africa in the Upper Cretaceous (Muller 1981).

Like Humiriaceae, there are "marginal" stomata on the disc and the anthers have a broad connective (Link 1992b); the wood anatomy is also similar.

Some information is taken from van Hooren and Nooteboom (1988b).

Centroplaceae + Bhesa: inflorescence branched, pedicels articulated; A 5, opposite K, 2 ovules/carpel, styles widely diverging, stigmas little expanded; exotegmic cells ribbon-shaped, thick-walled; embryo short.

CENTROPLACACEAE Doweld & Reveal   Back to Malpighiales

Tree; chemistry?; cork?; vessel member perforations?; sclereids +; stomata anisocytic; leaves 2-ranked, toothed, stipules cauline; plant dioecious; flowers small, disc? lobes alternating with K and outside A, staminate flowers: connective well developed, dehiscence oblique-apical, pollen psilate, pistillode +; carpellate flowers: C 0, ?staminodes minute, G [3], ovules subapical, epitropous; fruit a septicidal (and loculicidal) capsule opening from the base, K persistent; seed carunculate; exotesta rather tall, outer wall thickened, mesotegmic cells flattened and at right angles, endotegmen ± thick-walled; n = ?

1/1: Centroplacus glaucinus. W. Africa.

• Although hardly very distinctive, Centroplacaceae are dioecious evergreen trees with serrate and stipulate leaves that can be recognised by their branched inflorescences of small flowers; the staminate flowers have only five stamens in which the connective is well developed and the anthers dehisce by two oblique-apical slits.

Centroplacaceae may perhaps be sister to Ctenolophonaceae, diverging in the Cretaceous-Albian 111-100 million years before present ([109.6-]101.8[-96.6]/[97.1-]91.0[-88.1] million years before present: Davis et al. 2005a).

The endostome is lignified and more or less protruding.

Although Centroplacus glaucinus has often been placed in Pandaceae (Takhtajan 1997; Mabberley 1997), Webster (1994) and Radcliffe-Smith (2001) include it in Euphorbiaceae, but only with hesitation and with little certainty as to where it should be placed within the family. There is no obturator, unlike Euphorbiaceae. In a molecular study by Wurdack et al. (2004) Centroplacus is associated with Pandaceae, although with very little support, however, in Davis et al. (2005a) it is separate from Pandaceae and weakly associated with Ctenolophonaceae. Recognising the genus as a family seems most reasonable.

For information, see Forman (1966: general), Stuppy (1996: seed anatomy and good discussion, not Euphorbiaceae s.l.), Tokuoka and Tobe (2001: seed anatomy, Euphorbiaceae-Phyllanthoideae, but with some doubt), and Radcliffe-Smith (2001: generic description).

Bhesa   Back to Malpighiales

Trees; chemistry?; (cork mid-cortical); vessel elements with scalariform perforations; paratracheal parenchyma +; nodes 5:5; calcium oxalate as crystals [?always]; petiole with bundles forming a U or flattened-annular, 2-3 medullary bundles, (also wing bundles); leaves spiral, conduplicate, entire, petiole ± pulvinate apically, stipules almost encircling the stem, colleters +; inflorescence racemose; C contorted, A extrorse to introrse, nectary lobed or not, G [2], ovules basal, erect, apotropous, micropyle exostomal, outer integument 6-8 cells across, inner integument 4-5 cells across; K persistent; seed arillate, aril exostomal-funicular; exotegmic cells massive; n = ?; germination epigeal.

1/5. Indo-Malesia (Map: from Ding Hou 1962).

• Bhesa species are usually evergreen trees with stipulate leaves (the stipules almost surround the stem) the blades of which are entire and have strong, regularly ascending secondary veins and distinctive closely scalariform tertiary venation; there is a pulvinus at the apex of the petiole. The flowers are small and are rather undistinguished, but the fruits are strongly bilobed and have arillate seeds.

The ribbon-shaped exotegmic cells of Bhesa are longer than those of Centroplacus, and its integument is much thicker.

Bhesa was distinctive in morphological analyses of Celastraceae, in which it has up to now been included (Simmons & Hedin 1999; Matthews & Endress 2005b), if sometimes with some doubt (e.g. Pierre 1894 [he thought it might be in a separate family]; Metcalfe & Chalk 1950; Ding Hou 1962). Its distinct styles, scalariform perforation plates, etc., were somewhat out of place there, although Celastraceae were so heterogeneous that a strong case could not be made for its removal. The seed coat, with its massive exotegmic cells, is also very different, as is its pentalacunar nodes. Zhang and Simmons (2006) recently found that it fell among the few Malpighiales they included in their analysis of Celastrales, and Ken Wurdack (pers. comm.) suggests that a position around about here may be appropriate.

For general information, see Pierre (1894) and Ding Hou (1962), as Celastraceae. For seed and vegetative anatomy of B. ceylanica, see Jayasuriya & Balasubramaniam 3107.

HUMIRIACEAE Jussieu, nom. cons.   Back to Malpighiales

Trees; ellagic acid +; cork subepidermal; vessel elements with simple and scalariform perforation plates; true tracheids +; vestured pits +; nodes ?; sieve tube plastids with protein crystalloids and starch; mucilage cells frequent; stomata various; branching from previous flush; leaves often 2-ranked, involute (margins entire), tooth ?type, petiole short, stipules small or 0; inflorescence cymose; K connate, at least at base, A 10-many, filaments ± connate at least basally, forming a tube, anther sacs usu. 2, separated, superposed, connective broad, prolonged, exine usu. microreticulate, disc variable, from base of filaments to base of G, G [(4) 5(-7)], usu. opposite K, 1 (2) pendulous ovules/loculus, micropyle exo/(endo)stomal, style undivided, stigma slightly lobed, ?type; fruit a drupe, operculate, 1- or 2-seeded, surface sculpted; exotestal cells thick-walled, lignified, tegmen multiplicative [ca 5 cells thick], cross layer beneath exotegmen; endosperm copious, a little perisperm, embryo slightly curved; n = 12.

8[list]/50: Vantanea (16), Humiriastrum (12). Tropical America, W. Africa (Saccoglottis, also American) (Map: from Thorne 1973). [Photo - Flower] [Photo - Fruit]

• Humiriaceae are evergreen trees that may be recognised by their blackish-drying, short-petiolate, toothed leaves which elongate while still rolled up and often have longitudinal lines down the blade; the buds on growing shoots are long-pointed. The young twigs are angled. The cymose inflorescences have rather small flowers, and the anthers have conspicuously prolonged connectives and broadly rounded sepals. The fruit is a drupe and has a ridged, operculate stone that often has cavities containing almost resin-like material.

Humiriaceae may have diverged from other Malpighiales in the Cretaceous-Albian 111-100 million years before present ([117.2-]112.2[-108.9]/[102.8-]101.6[-100.3] million years before present: Davis et al. 2005a).

The fruits are dispersed by bats or by water, the empty cavities affording bouyancy.

Bove (1997) suggested that Ixonanthaceae were sister to Humiriaceae, both having ellagic acid, a "free" nectariferous disc encircling the ovary, and an undivided style with an entire stigma. Within Humiriaceae, Vantanea is sister to the other genera; it has three or more staminal whorls.

There is no endothelium.

Some information is taken from Boesewinkel (1985a); see Bove (1997) for a morphological phylogeny and Bove and Melhem (2000) for pollen.

PANDACEAE Engler & Gilg, nom. cons.   Back to Malpighiales

Trees to shrubs; cork?; vessel elements with (simple and - Galearia) scalariform perforations; rays 5-9 cells wide; sieve tubes with non-dispersive protein bodies; cristarque cells +; petiole bundles arcuate; cuticle waxes 0; leaves spiral and reduced on orthotropic axes, 2-ranked on plagiotropic axes, involute, a single vein running into the opaque persistent tooth apex; inflorescences various, plant dioecious, flowers small, K connate or free, C valvate or imbricate, nectary 0, staminate flowers: A = and opposite K, 10, or 15, in one or two series, basifixed, connective produced or not, pistillode +; carpellate flowers: staminodes 0, G [2-5], 1 pendulous epitropous ovule/carpel, micropyle endostomal, style +, (with styles), stigmas (bilobed), not expanded; fruit a drupe, surface often irregular; exotesta and endotegmen tanniniferous, exotegmen tracheoidal; endosperm ?development, +, cotyledons thin and flat, oily; n = 15.

3[list]/15: Microdesmis (10). Tropics, Africa to New Guinea (Map: approximate).

• Pandaceae are evergreen trees with stipulate leaves and often very strongly plagiotropic branches that have even been confused with compound leaves. The flowers are small and inconspicuous, the plants are dioecious, and the fruit is a drupe. [Photo - Habit] [Photo - Fruit].

Pandaceae seem to be a very old and isolated clade, dating back perhaps to the late Aptian (Cretaceous) 114-112 million years before present ([118.7-]113.8[-110.2]/[105.5-]101.6[-101.9] million years before present: Davis et al. 2005a).

Panda has straight [atropous] ovules; Microdesmis has punctate leaves. The plagiotropic branches have been confused with compound leaves, especially in the derived Galearia and Panda; the stipules may be asymmetrically placed, as in Panda.

Pandaceae are still often included in Euphorbiaceae, e.g. Govaerts et al. (2000) and Radcliffe-Smith (2001), but they differ from even the uniovulate taxa (Euphorbiaceae str.) in several respects, including their indehiscent fruits. Rays of Euphorbiaceae are only 1-5 cells wide (Hayden & Hayden 2000). Pandaceae lack obturators, Euphorbiaceae have them - another difference. Dicoelia (Euphorbiaceae - Dicoelieae) and Galearia both have stamens in depressions in the petals. Dicoelia has a low, thin-walled testa, a massive exotegmen, and a moderately thickened mesotegmen (Stuppy 1996), however, it is to be placed in Phyllanthaceae (Kathriarachchi et al. 2005). Centroplacus is also not included, although it is placed sister to Pandaceae, but without much support, by Wurdack et al. (2004); see Centroplacaceae here.

For information, see Forman (1966: general), Vaughan and Rest (1969), Hegnauer (1969: chemistry), Stuppy (1996: seed anatomy), Nowicke et al. (1998: pollen), Radcliffe-Smith (2001: generic descriptions) and Tokuoka and Tobe (2003: ovules and seeds [both integuments 3-5 cells across]). For a checklist and bibliography, see Govaerts et al. (2000, vol. 4).

Clusiaceae + Bonnetiaceae + Hypericaceae + Podostemaceae: xanthones common; vessel elements with simple perforations; nodes 1:1; stomata paracytic; leaves with colleters, stipules 0; inflorescence cymose, C contorted, A many, (fasciculate), nectary 0, G opposite K, or median member adaxial, many ovules/carpel, micropyle bistomal, stigma papillate; fruit a septicidal or -fragal capsule; exotegmen with anticlinal walls sinuous, low, lignified; endosperm at most slight, embryo ± fusiform.

This clade seems to have diverged in the Cretaceous-Albian, 111-100 million years before present, Clusiaceae (here sister to the rest of the clade) in turn diverging perhaps in the Cenomanian (104-)94(-92)/(95-)89(-87) million years before present (Davis et al. 2005a).

Morphological data in particular (most of the features in the characterisation above!) seemed to suggest a grouping of Elatinaceae + Bonneticaceae + Clusiaceae/Hypericaceae (e.g. see versions of this site prior to version 6). This was not a monophyletic group in Savolainen et al. (2000a), indeed, Ploiarium is there placed in Malvales, although testa anatomy, etc., are strongly against such a position. Although analyses in Chase et al. (2002) weakly link Elatinaceae and Bonnetiaceae + Clusiaceae + Podostemaceae, Elatinaceae are probably sister to Malpighiaceae (Davis & Chase 2004; Davis et al. 2005a; Tokuoka & Tobe 2006). Relationships within the Bonnetiaceae + Clusiaceae + Podostemaceae group remain unclear (see also Soltis et al. 1999b; Gustaffson et al. 2002; Davis et al. 2005b). Since Podostemaceae are strongly linked with Hypericum in particular, although the branch is rather long, I have tentatively split out Clusiaceae from Hypericaceae. Bonnetiaceae + Clusiaceae + Hypericaceae seem to be a distinct group with several potential synapomorphies (some are lost or highly modified in Podostemaceae), and they are recovered even in morphological analyses (e.g. Luna & Ochoterena 2004 - Hypericaceae not in the study).

CLUSIACEAE Lindley, nom. cons.//GUTTIFERAE Jussieu, nom. cons. et nom. alt.   Back to Malpighiales

Woody; flavones, flavonols, biflavonoids, (ellagic acid) +; (vessel elements with scalariform perforations); schizogenous canals or cavities +; leaves opposite, often flat (conduplicate), margins entire, with gland dots or lines, (paired "stipular glands" on the stem; colleters 0); flowers 4-5-merous, C (contorted), (0-)4-5(-8), A (5-)many, connate or not, anthers often with glands, G (?1) [2-5(-12)], (placentation apical, basal, parietal, 1-few ovules/carpel), stigmas much expanded to punctate, not papillate, wet; (fruit a berry or drupe); seeds 1-many, (exotegmen 0, then testa usu. much developed); embryo green or white.

27[list]/1050 - two subfamilies below. Pan tropical.

1. Kielmeyeroideae Engler

Leaves (spiral, 2-ranked; supervolute - Kielmeyera), with pellucid dots (canals); androecium rarely obviously fasciculate, anthers glands 0 to small, style (styles) +, usually long; cotyledons moderate sized to huge [Calophyllum, Mesua, etc.].

13/460: Calophyllum (190), Kayea (70), Mammea (70), Kielmeyera (50). Throughout the tropics (Map: in part see Stevens 1980 - blue is Calophyllum inophyllum). [Photo - Flower] [Photo - Flower]

Recent work suggest that Clusiella is to be included in this clade (Gustaffson et al. 2002). Its seeds and vegetative anatomy (including that of the root) are consistent with this position, although the flowers are a little odd, since they do indeed look like those of Clusia...

Synonymy: Calophyllaceae J. Agardh

2. Clusioideae Engler

(Root cork cambium superficial in origin - Clusia, etc.); leaves usu. with canals; (plant dioecious - Garcinia, etc.); androecium obviously fasciculate or not, anthers glands usu. 0, style or separate styles +, usually short; (seed arillate); hypocotyl much enlarged, cotyledons minute.

14/595: Garcinia (240: esp. Old World), Clusia (300-400: entirely American), Chrysochlamys (55). Throughout the tropics (Map: incomplete, from Gustaffson et al. 2007). [Photo - Staminate flower, Fruit].

Roots of Clusia have superficial phellogen, fairly common in epiphytic taxa in general.

Synonymy: Cambogiaceae Horaninow, Garciniaceae Bartling

• Clusiaceae may be recognised by their usually opposite, exstipulate leaves with entire margins and usually quite obvious glands or canals; there is often some exudate from the cut twig. The flowers have free sepals and petals and numerous stamens, quite often in fascicles; the anthers may bear various kinds of glands. The fruit, when dehiscent, opens along the septal radius.

• Kielmeyeroideae are usually trees. Their leaves have pellucid dots, rarely canals, and the androecium is not obviously fasciculate, the anthers sometimes having large, apical glands. The seeds range from small to large; when large, they are made up almost entirely of two huge cotyledons.

• Clusioideae range from shrubs or trees, some are epiphytic; many are dioecious. The leaves usually have canals and the androecium is sometimes obviously fasciculate. The seeds are small to large, arils being known from many New World taxa, and the embryos are often apparently undifferentiated - the cotyledons are small to almost absent and the hypocotyl is enormous.

For discussion on an interesting and well-preserved late-Cretaceous fossil from ca 90 million years before present and possibly assignable to Clusioideae, see Crepet & Nixon (1998). Variation in the androecium and gynoecium in Clusia and Garcinia is extreme. In the former genus, resins (polyisoprenylated benzophenones, mixed with fatty acids) are quite commonly a floral reward, and floral resin production may have evolved four times (and in Clusiella: Gustaffson & Bittrich 2002). Reins are an uncommon floral reward (but see also Dalechampia - Euphorbiaceae). Clusia is a very diverse genus, including epiphytes and stanglers, and is also diverse at elevations up to 3500 m in altitude (Gustafsson et al. 2007; for the general ecology of Clusia, see papers in Lüttge 2007).

Marila asymmetralis, alone in the family, has obliquely monosymmetric flowers. Alternate-leaved genera form a clade in Kielmeyeroideae; these genera also have capsular fruits, often with winged seeds, and their embryos have cotyledons with cordate bases. Many Theaceae have capsular fruits, winged seeds, and flowers with many stamens, and these alternate-leaved Clusiaceae used to be placed in that family... Another pseudoproblem caused by "intermediates" between groups which turn out not to be closely related at all (cf. Baretta-Kuipers 1976).

For a preliminary phylogeny, see Gustafsson et al. (2002).

For chemistry, see Hegnauer (1966, 1989, as Guttiferae), for general information, see Stevens (2006c).

BONNETIACEAE Nakai   Back to Malpighiales

Evergreen shrubs; plant glabrous; (nodes 3:3); mucilage cells common; leaves spiral, involute [Keller 1996] or supervolute, margins minutely toothed by setae, petiole short; (A fasciculate; nectary fasciclodial), G [3-5], micropyle exostomal, style or separate styles +, stigma papillate; cotyledons usu. small (-50% the embryo); n = ca 150 [Bonnetia cubensis].

3[list]/35: Bonnetia (30). Cambodia, Malesia (mostly Western), Cuba, South America. [Photo - Flower] [Photo - Flower]

• Bonnetiaceae may be recognised by their long-pointed buds, usually rather closely-set, spiral, short-petiolate and minutely-serrate leaves. The moderate-sized flowers are borne in cymose inflorescences, the corolla is contorted, the stamens are numerous, and the septicidal capsules contain many small seeds.

For Archytaea, Wawra de Fernsee (1886) shows a floral diagram in which both the five carpels and the stamen fascicles are drawn opposite the calyx. Bonnettia s.l. has trilacunar nodes, a mucilaginous epidermis, a foliar endodermis, and foliar sclereids; Archytaea and Ploiarium have unilacunar nodes and lack the distinctive epidermis, foliar endodermis and sclereids (Dickison & Weitzmann 1996). Takhtajan (1993) describes the pith as having secretory canals, as in Clusiaceae (cf. Baretta-Kuipers 1976).

Ovule morphology in Bonnetiaceae appears to be unknown.

For chemistry, see Hegnauer (19690, as Theaceaee), and for a general account, see Weitzman et al. (2006).

Hypericaceae + Podostemaceae: ovules tenuinucellate.

Divergence between Podostemaceae and Hypericaceae may have occured in the Campanian, (82-)76(-69)/(74-)72(-66) million years before present (Davis et al. 2005a).

HYPERICACEAE Jussieu, nom. cons.   Back to Malpighiales

Shrubs or annual to perennial herbs; flavones, flavonols, biflavonoids, (ellagic acid) +; cork pericyclic; polyderm widespread; schizogenous cavities +; leaves opposite, margins entire, with gland dots or lines; flowers 4-5-merous, C often contorted, A (5-)many, often fasciculate, centrifugal, anthers often with small glands, G [3-5], (placentation parietal), style + to styles separate, (stigma not papillate, dry; (fruit a berry or drupe); seeds (5-)many, (exotegmen 0); embryo green or white, cotyledons moderate-sized (-80% of the embryo); n = 6-12, etc.

9[list]/560: Hypericum (370), Vismia (55), Harungana (50). World-wide (map: from Hultén & Fries 1986; Meusel et al. 1978 - still incomplete for Africa). [Photo - Flower]

• Hypericaceae include numerous herbs or shrubs. Their leaves are opposite and with pellucid or dark dots; the androecium is often fasciculate; the stigma is sometimes obviously papillate; and the seeds are small (usually less than 4 mm long).

For chemistry, see Hegnauer (1966, 1989, as Guttiferae), for androecial development, see Leins (2000), and for general information, see Stevens (2006c).

Synonymy: Ascyraceae Plenck

PODOSTEMACEAE Kunth, nom. cons.   Back to Malpighiales

Annual (perennial) herbs of fast-flowing water, plant ± thalloid, stem root and leaf often not distinguishable, plant attached to substrate by haptera; xanthones +; primary root 0, other roots dorsiventrally flattened, photosynthetic, exogenous or endogenous, shoots endogenous [also at least sometimes flowers], branching extra-axillary; cork?; vessels usu 0; latex/resin cells canals +; epidermal SiO₂ bodies common; cuticle waxes 0; when leaves present spiral, opposite, 2- or 3-ranked, leaf base broad or not, stipules petiolar or 0; flowers also solitary, monosymmetric; P in single whorl, A 1-many (2 whorls, inner extrorse), filaments often basally connate (connective prolonged), pollen microechinate, infratectum granular, when G equalling P, opposite to it, (2-few ovules/carpel), micropyle exostomal, embryo sac monosporic [from the subchalazal spore], tetranucleate [Apinagia type], no polar nuclei or double fertilisation, nucellus plasmodial, style + (0), stigma linear; capsule ribbed, about the same size as the ovary, pedicels elongating; exotesta thick-walled, often mucilaginous, (exo- and) endotegmen lignified; suspensor [micropylar] haustorium +; cotyledons large; n = 10.

48[list]/270 - three subfamilies below. Usually tropical, esp. America, many genera are monotypic.

1. Tristichoideae (J. C. Willis) Engler

Xanthones?; stomata?; P 3, (connate), A (1-)3, anthers sagittate, pollen pantoporate, G [3], integuments develop simultaneously; hypocotyl 0.

3/4-10. India-Australia, but Tristicha trifaria in Africa and America (Map: from van Royen 1953; Cusset & Cusset 1988a; Kito & Kato 2004).

Some species of Dalzellia have a cupule at the base of the pedicel that is formed by leafy shoot axes (Mathew et al. 2001). The inner and outer integuments develop together and the nucellus is coenocytic. Distribution of these characters? Tristicha - A 1, adaxial, median carpel abaxial? (Schnell 1998).

Synonymy: Philocrenaceae Bongard, Tristichaceae J. C. Willis

Weddellinoideae + Podostemoideae: G [2], with apical septum.

2. Weddellinoideae (Warming) Engler

Flowers single, terminal; P (4) 5 (6), 1-veined, A 5-25, anthers X-shaped, pollen ?development, smooth, ?infratectum?, integuments develop simultaneously, stigmas globose; capsule not ribbed; tegmen (?layer) thick walled; hypocotyl +.

1/1: Weddellina squamulosa. N. South America (Map: from van Royen 1953).

3. Podostemoideae (Warming) Engler

Shoot apical meristem 0 [whole plant "leafy"]; apical meristem of root on the underside, (roots exogenous); ("laticiferous" tubes +); stomata 0?; leaves often distichous, ensiform, but bifacial, some leaves dithecous [double-sheathed, one sheath on both sides]; flowers or groups of flowers enveloped by a non-vascularized spathella [= bracteoles?]; P 2-25, often 2-3 on one side, lobes narrow, sometimes replaced by stamens, A 1-3(-many), often sagittate, (extrorse), (microsporogenesis successive [tetrads tetragonal]), pollen often in diads, 3-5-colpate, G also [3(-7)], (unilocular), outer integument develops first, nucellus plasmodial before fertilisation, (embryo sac bisporic [Polypleurum and Podostemon types]), style short, branches long; hypocotyl 0, radicle 0; (n = 14). Floral Diagram.

45/260: Apinagia (50: perhaps paraphyletic, see Philbrick et al. 2001). Pantropical (Map: from van Royen 1951; van Steenis 1972). [Photo - Marathrum Flower]

There is sometimes an apical meristem in the vegetative body and the spathella may be produced by the connation of two foliar structures (Jäger-Zürn 2005b). The seedling of Zeylanidium olivaceum has a hypocotyl.

Synonymy: Marathraceae Dumortier

• Podostemaceae are the quintessential aquatic angiosperm with a vegetative body so highly modified that conventional terms cannot be used to describe it - the plant body of many taxa is best described as thalloid or fucoid, although some species look like a demented apiaceous plant with 2-ranked leaves. Many species are annual, and flowers or groups of flowers are often enveloped in a spathella; the capsule is septicidal, and the small seeds are often mucilaginous.

Although there have been suggestions that Podostemaceae attached to rocks by means of a special glue that they produced, it is more likely that it is materials in a biofilm produced by associated cyanobacteria that attach the plant to the substrate. There are hooked hairs on the lower side of the thallus that stick to the cyanobacterial filaments and associated biofilm. Indeed, these cyanobacteria may even produce nitrogen used by the plant; Podostemaceae usually grow in oligotrophic rivers flowing over gneiss or granite, being absent in rivers ove